Thermal percolation in polymer nanocomposites─the rapid increase in thermal transport due to the formation of networks among fillers─is the subject of great interest in thermal management ranging from general utility in multifunctional nanocomposites to high-conductivity applications such as thermal interface materials. However, It remains a challenging subject encompassing both experimental and modeling hurdles. Successful reports of thermal percolation are exclusively found in high-aspect-ratio, conductive fillers such as graphene, albeit at filler loadings significantly higher than the electrical percolation threshold. This anomaly was attributed to the lower filler–matrix thermal conductivity contrast ratio kf/km ∼104 compared to electrical conductivity ∼1012–1016. In a randomly dispersed composite, the effect of a low contrast ratio is further accentuated by uncertainties in the morphology of the percolating network and presence of other phases such as disconnected aggregates and colloidal dispersions. Thus, the general properties of percolating networks are convoluted as they lack a defined structure. In contrast, a prototypical system with controllable nanofiller placement enables the elucidation of structure–property relations such as filler size, loading, and assembly. Using self-assembled nanocomposites with a controlled 1,2,3-dimension nanoparticle (NP) arrangement, we demonstrate that thermal percolation can be achieved in spite of using spherical, nonconductive fillers (kf/km ∼60) at a low volume fraction (9 vol %). We observe that the effects of volume fraction, interfacial thermal resistance, and filler conductivity on thermal conductivity depart from effective medium approximations. Most notably, contrast ratio plays a minor role in thermal percolation above kf/km ∼60─a common range for semiconducting nanoparticles/polymer ratios. Our findings bring new perspectives and insights to thermal percolation in nanocomposites, where the limits in contrast ratio, interfacial thermal conductance, and filler size are established.
Abstract:
Publication date:
March 21, 2022
Publication type:
Journal Article